Bilateral loss of vestibular sensation due to ototoxic injury or other insults to both labyrinths is disabling. Affected individuals suffer chronic disequilibrium, increased risk of falls and unstable vision during head movements typical of common daily activities like walking and driving. Most individuals with mild or moderate loss eventually compensate through rehabilitative exercises that augment residual function, but those with profound loss who fail to compensate have no good treatment options. The vestibular nerves are intact in many such cases, so an implanted stimulator encoding signals from a head-mounted motion sensor can excite the vestibular nerve and - if it creates the right patterns of activity on the nerve's five branches - restore sensation of head movement (much as a cochlear implant restores sensation of sound). This proposal builds on substantial progress we have already made toward that goal, including: (1) characterization of the three dimensional vestibulo-ocular reflex (3D VOR), vestibular nerve activity, and inner ear histology in animals with ototoxic injury after gentamicin treatment; (2) development of a multi-channel vestibular prosthesis (MVP) that encodes head rotation via electrical stimulation of the three ampullary nerve branches innervating the semicircular canals (SCCs); (3) demonstration that the MVP can significantly restore the 3D VOR in rodents and rhesus monkeys with bilateral vestibular deficiency (BVD) while preserving hearing; (4) development of powerful computational strategies for minimizing the difference between MVP-evoked and ideal 3D VOR responses; and (5) confirmation that the central nervous system (CNS) rapidly adapts to correct much of the remaining error. While generating these promising results, we identified three major remaining challenges: (1) VOR misalignment (due to current spread activating nontarget nerve fibers outside the targeted branch of the vestibular nerve), (2) VOR asymmetry (due to the MVP's inability to suppress spontaneous neuronal activity when trying to encode inhibitory head movements) and (3) lack of utricle and saccule input (because we have focused effort on stimulating the ampullary nerves to restore SCC and angular VOR function and have not yet attempted prosthetic stimulation of the macular [utricular and saccular] nerves to restore tilt and translational VOR reflexes they normally drive). In this project, we will study ways to overcome these challenges using a unique combination of techniques including binocular 3D VOR measurements, single- unit recording in alert rodents and rhesus monkeys, a new MVP that encodes both rotational and translational movement, a powerful computational model of the implanted labyrinth, electrically-evoked compound action potentials (eCAPs), a novel rehabilitation paradigm, and a highly innovative method for controlling neuronal activity. These studies will yield new electrode designs, stimulus optimization protocols, computer-aided design tools, surgical techniques and rehab paradigms that set the stage for successful deployment of MVPs for clinical treatment of tens of thousands of individuals chronically disabled by loss of vestibular sensation.